Cytokines play a central role in the communication between cells. Secretion of a cytokine from one cell in response to a stimulus can trigger an adjacent cell to undergo an appropriate biological response—such as stimulation, differentiation, or apoptosis. It is hypothesized that important biological events can be influenced not only by affecting cytokine release from the first cell, but also by binding to receptors on the second cell, which mediates the subsequent response. The invention described in this patent application provides new compounds for affecting signal transduction from tumor necrosis factor.
The cytokine known as tumor necrosis factor (TNF or TNF-α) is structurally related to lymphotoxin (LT or TNF-β). They have about 40 percent amino acid sequence homology (Old, Nature 330:602-603, 1987). These cytokines are released by macrophages, monocytes and natural killer cells and play a role in inflammatory and immunological events. The two cytokines cause a broad spectrum of effects both in vitro and in vivo, including: (i) vascular thrombosis and tumor necrosis; (ii) inflammation; (iii) activation of macrophages and neutrophils; (iv) leukdcytosis; (v) apoptosis; and (vi) shock. TNF has been associated with a variety of disease states including various forms of cancer, arthritis, psoriasis, endotoxic shock, sepsis, autoimmune diseases, infections, obesity, and cachexia. TNF appears to play a role in the three factors contributing to body weight control: intake, expenditure, and storage of energy (Rothwell, Int. J. Obesity 17:S98-S101, 1993). In septicemia, increased endotoxin concentrations appear to raise TNF levels (Beutler et al. Science 229:869-871, 1985).
Attempts have been made to alter the course of a disease by treating the patient with TNF inhibitors, with varying degrees of success. For example, the TNF inhibitor dexanabinol provided protection against TNF mediated effects following traumatic brain injury (Shohami et al. J. Neuroimmun. 72:169-77, 1997). Some improvement in Crohn's disease was afforded by treatment with anti-TNF antibodies (Neurath et al., Eur. J. Immun. 27:1743-50, 1997).
Human TNF and LT mediate their biological activities by binding specifically to two distinct glycoprotein plasma membrane receptors (55 kDa and 75 kDa in size, known as p55 and p75 TNF-R, respectively). The two receptors share 28 percent amino acid sequence homology in their extracellular domains, which are composed of four repeating cysteine-rich regions (Tartaglia and Goeddel, Immunol. Today 13:151-153, 1992). However, the receptors lack significant sequence homology in their intracellular domains, and mediate different intracellular responses to receptor activation. In accordance with the different activities of TNF and LT, most human cells express low levels of both TNF receptors: about 2,000 to 10,000 receptors per cell (Brockhaus et al., Proc. Natl. Acad. Sci. USA 87:3127-3131, 1990).
Expression of TNF receptors on both lymphoid and non-lymphoid cells can be influenced experimentally by many different agents, such as bacterial lipopolysaccharide (LPS), phorbol myristate acetate (PMA; a protein kinase C activator), interleukin-1 (IL-1), interferon-gamma (IFN-γ) and IL-2 (Gatanaga et al. Cell Immunol. 138:1-10, 1991; Yui et al. Placenta 15:819-835, 1994). It has been shown that complexes of human TNF bound to its receptor are internalized from the cell membrane, and then the receptor is either degraded or recycled (Armitage, Curr. Opin. Immunol. 6:407-413, 1994). It has been proposed that TNF receptor activity can be modulated using peptides that bind intracellularly to the receptor, or which bind to the ligand binding site, or that affect receptor shedding. See for example patent publications WO 95/31544, WO 95/33051, WO 96/01642, and EP 568 925.
TNF binding proteins (TNF-BP) have been identified at elevated levels in the serum and urine of febrile patients, patients with renal failure, and cancer patients, and even certain healthy individuals. Human brain and ovarian tumors produced high serum levels of TNF-BP. These molecules have been purified, characterized, and cloned (Gatanaga et al., Lymphokine Res. 9:225-229, 1990a; Gatanaga et al., Proc. Natl. Acad. Sci USA 87:8781-8784, 1990b). Human TNF-BP consists of 30 kDa and 40 kDa proteins which are identical to the N-terminal extracellular domains of p55 and p75 TNF receptors, respectively (U.S. Pat. No. 5,395,760; EP 418,014). Such proteins have been suggested for use in treating endotoxic shock. Mohler et al. J. Immunol. 151:1548-1561, 1993.
There are several mechanisms possible for the production of secreted proteins resembling membrane bound receptors. One involves translation from alternatively spliced mRNAs lacking transmembrane and cytoplasmic regions. Another involves proteolytic cleavage of the intact membrane receptors, followed by shedding of the cleaved receptor from the cell. The soluble form of p55 and p75 TNF-R do not appear to be generated from mRNA splicing, since only full length receptor mRNA has been detected in human cells in vitro (Gatanaga et al., 1991). Carboxyl-terminal sequencing and mutation studies on human p55 TNF-R indicates that a cleavage site may exist between residues Asn 172 and Val 173 (Gullberg et al. Eur. J. Cell. Biol. 58:307-312, 1992).
There are reports that a specific metalloprotease inhibitor, TNF-α protease inhibitor (TAPI) blocks the shedding of soluble p75 and p55 TNF-R (Crowe et al. J. Exp. Med. 181:1205-1210, 1995; Mullberg et al. J. Immunol. 155:5198-5205, 1995). The processing of pro-TNF on the cell membrane to release the TNF ligand appears to be dependent on a matrix metalloprotease like enzyme (Gearing et al. Nature 370:555-557, 1994). This is a family of structurally related matrix-degrading enzymes that play a major role in tissue remodeling and repair associated with development and inflammation (Birkedal-Hansen et al. Crit. Rev. Oral Biol. Med. 4:197-250, 1993). The enzymes have Zn2+ in their catalytic domains, and Ca2+ stabilizes their tertiary structure significantly.
In European patent application EP 657536A1, Wallach et al. suggest that it would be possible to obtain an enzyme that cleaves the 55,000 kDa TNF receptor by finding a mutated form of the receptor that is not cleaved by the enzyme, but still binds to it. The only proposed source for the enzyme is a detergent extract of membranes for cells that appear to have the protease activity. If it were possible to obtain an enzyme according to this scheme, then the enzyme would presumably comprise a membrane spanning region. The patent application does not describe any protease that was actually obtained.
In a previous patent application in the present series (International Patent Publication WO 9820140), methods are described for obtaining an isolated enzyme that cleaves both the p55 and p75 TNF-R from cell surfaces. A convenient source is the culture medium of cells that have been stimulated with phorbol myristate acetate (PMA). The enzyme activity was given the name TRRE (TNF receptor releasing enzyme). In other studies, TRRE was released immediately upon PMA stimulation, indicating that it is presynthesized in an inactive form to be rapidly converted to the active form upon stimulation. Evidence for direct cleavage of TNF-R is that the shedding begins very quickly (˜5 min) with maximal shedding within 30 min. TRRE is specific for the TNF-R, and does not cleave IL-1 receptors, CD30, ICAM-1 or CD11b. TRRE activity is enhanced by adding Ca++ or Zn++, and inhibited by EDTA and phenaritroline.
Given the involvement of TNF in a variety of pathological conditions, it is desirable to obtain a variety of factors that would allow receptor shedding to be modulated, thereby controlling the signal transduction from TNF at a disease site.